Birth defects, which occur in one out of 20 live births, often affect multiple organs that have common developmental origins. Human and mouse studies indicate that haploinsufficiency of the transcription factor TBX1 disrupts pharyngeal arch development, resulting in the cardiac and craniofacial features associated with microdeletion of 22q11 (del22q11), the most frequent human deletion syndrome. Here, we have generated an allelic series of Tbx1 deficiency that reveals a lower critical threshold for Tbx1 activity in the cardiac outflow tract compared with other pharyngeal arch derivatives, including the palatal bones. Mice hypomorphic for Tbx1 failed to activate expression of the forkhead transcription factor Foxa2 in the pharyngeal mesoderm, which contains cardiac outflow precursors derived from the anterior heart field. We identified a Fox-binding site upstream of Tbx1 that interacted with Foxa2 and was necessary for pharyngeal mesoderm expression of Tbx1, revealing an autoregulatory loop that may explain the increased cardiac sensitivity to Tbx1 dose. Downstream of Tbx1, we found a fibroblast growth factor 8 (Fgf8) enhancer that was dependent on Tbx1 in vivo for regulating expression in the cardiac outflow tract, but not in pharyngeal arches. Consistent with its role in regulating cardiac outflow tract cells Tbx1 gain of function resulted in expansion of the cardiac outflow tract segment derived from the anterior heart field as marked by Fgf10. These findings reveal a Tbx1-dependent transcriptional and signaling network in the cardiac outflow tract that renders mouse cardiovascular development more susceptible than craniofacial development to a reduction in Tbx1 dose, similar to humans with del22q11.

Haploinsufficiency of Tbx1 is likely a major determinant of cardiac and craniofacial birth defects associated with DiGeorge syndrome. Although mice deficient in Tbx1 exhibit pharyngeal and aortic arch defects, the developmental program and mechanisms through which Tbx1 functions are relatively unknown. We identified a single cis-element upstream of Tbx1 that recognized winged helix/forkhead box (Fox)-containing transcription factors and was essential for regulation of Tbx1 transcription in the pharyngeal endoderm and head mesenchyme. The Tbx1 regulatory region was responsive to signaling by Sonic hedgehog (Shh) in vivo. We show that Shh is necessary for aortic arch development, similar to Tbx1, and is also required for expression of Foxa2 and Foxc2 in the pharyngeal endoderm and head mesenchyme, respectively. Foxa2, Foxc1, or Foxc2 could bind and activate transcription through the critical cis-element upstream of Tbx1, and Foxc proteins were required, within their expression domains, for Tbx1 transcription in vivo. We propose that Tbx1 is a direct transcriptional target of Fox proteins and that Fox proteins may serve an intermediary role in Shh regulation of Tbx1.

Appropriate interactions between the epithelium and adjacent neural crest-derived mesenchyme are necessary for normal pharyngeal arch development. Disruption of pharyngeal arch development in humans underlies many of the craniofacial defects observed in the 22q11.2 deletion syndrome (del22q11), but the genes responsible remain unknown. Tbx1 is a T-box transcription factor that lies in the 22q11.2 locus. Tbx1 transcripts were found to be localized to the pharyngeal endoderm and the mesodermal core of the pharyngeal arches, but were not present in the neural crest-derived mesenchyme of the pharyngeal arches. Sonic hedgehog (Shh) is also expressed in the pharyngeal arches and is necessary for normal craniofacial development. We found that Tbx1 expression was dependent upon Shh signaling in mouse embryos, consistent with their overlapping expression in the pharyngeal arches. Furthermore, Shh was sufficient to induce Tbx1 expression when misexpressed in selected regions of chick embryos. These studies reveal a Shh-mediated pathway that regulates Tbx1 during pharyngeal arch development.

Nkx2.5/Csx and dHAND/Hand2 are conserved transcription factors that are coexpressed in the precardiac mesoderm and early heart tube and control distinct developmental events during cardiogenesis. To understand whether Nkx2.5 and dHAND may function in overlapping genetic pathways, we generated mouse embryos lacking both Nkx2.5 and dHAND. Mice heterozygous for mutant alleles of Nkx2.5 and dHAND were viable. Although single Nkx2.5 or dHAND mutants have a morphological atrial and single ventricular chamber, Nkx2.5(-/-)dHAND(-/-) mutants had only a single cardiac chamber which was molecularly defined as the atrium. Complete ventricular dysgenesis was observed in Nkx2.5(-/-)dHAND(-/-) mutants; however, a precursor pool of ventricular cardiomyocytes was identified on the ventral surface of the heart tube. Because Nkx2.5 mutants failed to activate eHAND expression even in the early precardiac mesoderm, the Nkx2.5(-/-)dHAND(-/-) phenotype appears to reflect an effectively null state of dHAND and eHAND. Cell fate analysis in dHAND mutants suggests a role of HAND genes in survival and expansion of the ventricular segment, but not in specification of ventricular cardiomyocytes. Our molecular analyses also revealed the cooperative regulation of the homeodomain protein, Irx4, by Nkx2.5 and dHAND. These studies provide the first demonstration of gene mutations that result in ablation of the entire ventricular segment of the mammalian heart, and reveal essential transcriptional pathways for ventricular formation.